I've seen a lot of startups that did plenty of market research (including both my last failed startup and my last employer), but I've yet to see one get anything useful out of it.

There're a couple of pitfalls to be wary of when doing market research, like:

1.) Confirmation bias. At my last employer, we had an intern spend all summer calling prospective customers. He called 30-40 organizations and got through to a grand total of 2 executives, who were sorta lukewarm about the idea. Then my boss said, "Well, I think there's enough interest to move forwards with this project." Then we busted our ass for a year, put it in front of a bunch of other customers, incorporated their feedback, and it still didn't sell.

2.) Not having something for people to react to. People don't know what they want, but if you show them something, they can often tell you whether they like it or not, and possibly suggest things they do want.

3.) Calling the wrong people. It doesn't help to get the input of people who won't use your product anyway. This can be kinda tricky though, since you often don't know who'll end up using your product ahead of time.